The Cyberspace Administration of China (CAC) has initiated an investigation into AI companies DeepSeek and Moonshot AI. The probe follows a September report from Anthropic, which alleged that seven Chinese developers used its Claude model for unauthorized model distillation.
According to Anthropic, Moonshot redirected over 23 million user queries to Claude between May and July, while DeepSeek sent more than 12 million queries over a 14-day period in July. Anthropic claims these queries contained sensitive data from government entities, police departments, and private organizations.
The investigation involves interviews with executives and staff at the targeted firms. Chinese authorities are evaluating whether these actions violated national regulations regarding the cross-border transfer of sensitive data, which is strictly controlled by the government.
The incident has gained political significance ahead of discussions between Chinese and US leadership regarding an AI dialogue mechanism. This framework aims to establish notification procedures for AI-related security incidents that could impact national interests.
